CN103646404A - Color-based rapid tea flower segmenting and counting method - Google Patents

Abstract

The invention relates to the technical field of tea and image processing and discloses a color-based rapid tea flower segmenting and counting method. The method comprises the following steps: converting an original tea RGB (Red, Green and Blue) colored image into an HSI (Hue, Saturation and Intensity) color space, performing image enhancement processing, performing feature tone-based convergence calculation on the tone H, and converting the image back to an RGB color space; performing preliminary seed selection according to the R and G parameters of the tea flower by using an improved fast area growth and merge algorithm, performing area growth on the seed area based on the color similarity and area adjacency, performing area growth and merge by combining the color distance and merge rules, and finally segmenting and counting the tea flower. The experimental result proves that the algorithm has high connectivity, and multiple tea flowers can be simply and rapidly segmented from tea images.

Description

A kind of quick Tea Flower based on color is cut apart and method of counting

Technical field

The invention belongs to tealeaves and technical field of image processing, is a kind of method that strengthens and cut apart and count in conjunction with the Tea Flower in tea tree of improved region growing and merge algorithm based on color.

Background technology

The development and application of Tea Flower is a popular direction in recent years, in Tea Flower, containing composition, there is removing toxic substances, antibacterial, hypoglycemic, delay senility and strengthen the effects such as immunity, its protein, tea polysaccharide, Tea Polyphenols, active polyphenoils exceed allied substances content in tealeaves, and residues of pesticides and content of beary metal are very low.

One strain tea tree, can grow nearly thousand vegetative buds of cooking tealeaves in general 1 year, more more reproduction bud raising up seed of also growing simultaneously.Along with applying of vegetative propagation technique, reproduction bud is no longer undertaken raising up seed of task, a large amount of Tea Flowers has become with tea tree germ and Ye Zhengshui strives fertile burden, in order to guarantee quality and the quantity of tealeaves in the coming year, tea grower will take the method for artificial pruning and spray plant growth regulator to suppress the breeding growth of Tea Flower every year.

Now Tea Flower fresh flower or dried flower being made to magma and composite powder by deep processing, can insert in the products such as food, beverage, daily cosmetics, women and children's amenities, have good function, is a kind of rare natural compounding type raw material.Utilize receive flower winter slack simultaneously, can increase agricultural incomes, also can increase tea yield in the coming year.

For colored image, main partitioning algorithm has edge detection method, clustering method, the method based on region etc., edge detection method is for coloured image object and the object in positioning image exactly often, and clustering method calculated amount is too large, region-growing method can directly act on color space, in algorithm, can take into full account that image color distributes and the connectivity of region etc., thus of the present invention in the cutting apart and count of Tea Flower, selective area growth method.But algorithm of region growing is often subject to the impact of choosing of initial seed point and succession, also has definite area growth and merge regular problem.

The present invention is directed to the color characteristic of Tea Flower coloured image in tealeaves image, proposed a kind of simple and quick Tea Flower based on color and cut apart and method of counting.First the method passes through the conversion in color of image space by original RGB tealeaves image, in HSI color space, strengthened the feature of Tea Flower in H tone, and in cutting apart subsequently, adopted improved algorithm of region growing, the method has been improved wherein choosing method and growth rule and the merging rule of Seed Points on the basis of traditional area growth algorithm.The results show, the method is not only simple efficient, and when guaranteeing the connectivity of region, can be consistent with human eye vision.

Summary of the invention

The initial coloured image of a common width is to be all created on RGB color space, by R, G, tri-representation in components of B.RGB color component shows good effect for color, but general and be not suitable for color analysis, because have very large relevance between R, G, tri-components of B.But for Tea Flower image, but because Tea Flower is white, take, green and bottle green are very outstanding in main tealeaves image, and the object of the invention is fast Tea Flower to be detected, to facilitate the suitable time to pluck, experiment discovery, the R of Tea Flower, two color components of G are obviously different from the true qualities of tealeaves and tea shoot.So in the present invention, first original tealeaves RGB coloured image is carried out to the conversion of color space, by image, from RGB color space conversion, be HSI color space, and the feature tone H of Tea Flower is wherein carried out to the enhancing that image converges and process, to reduce the level of color, facilitate region growing and the merging in later stage, thereafter color of image space is converted back to RGB color space again; Then the fast area of application enhancements is grown and merge algorithm, according to the R of Tea Flower, G parameter, carrying out preliminary seed selects, similarity to seed region based on color and the adjacency in region are carried out region growing, and color combining distance and merge rule and carry out final growth and the merging in region, finally complete the Fast Segmentation of Tea Flower and counting.

A kind of quick Tea Flower based on color of the present invention is cut apart and method of counting, and the main process of its algorithm is:

(1) obtain the original image of tealeaves;

(2) by original image from RGB color space conversion to HSI color space, and the figure image intensifying that tone H wherein carries out based on feature tone is converged to calculating;

(3) image is converted back to RGB color space;

(4), according to the R of Tea Flower, G parameter, in image, select Partial Feature pixel as seed;

(5) based on growth rule, seed region is grown, the neighbor similar to seed color character is attached on the seed of growth district;

(6) based on merging rule and color distance, a plurality of sub-blocks region of entire image is scanned and chosen, to close in color, region adjacent on space merges;

(7) expand and the morphology that shrinks is processed in the region after being combined;

(8) complete cutting apart and counting of a plurality of Tea Flowers.

H span in HSI color space is [0, 360], and in tealeaves image, Tea Flower has outstanding white, pistil has outstanding yellow, the tone images Enhancement Method that the present invention proposes is: defined two feature TINTCs of tea tree petal and pistil, in hue circle plane, experiment finds that these two tone value correspondences are respectively 30 and at 43 o'clock, can obtain good effect, so centered by these two values, near the tone value central value satisfying condition is converged to central value by different conditions and step-length, to reduce the number of relatively less important tone, make each tone value closer to center tone, reach and make the stronger object of original image gradation sense, also be more conducive to cutting apart of next step Tea Flower simultaneously.

Step-length is very important to choosing of converging, and step-length is crossed conference and caused image to occur the ill effects such as color change, and the too small meeting of step-length makes regulating effect not obvious.Through many experiments test, determine the step-length that converges that step-length elects 1,3 as, 5 three kind is different, little, more larger away from the step-length of TINTC near the step-length of TINTC.The calculating of converging of tone is defined as:

H _i′＝H _i±k

1＜|H _i-H ₀|≤5 k＝±1

Wherein, 5 < | H _i-H ₀|≤10 k=± 3

10＜|H _i-H ₀|≤15 k＝±5

H _ifor the tone value of certain point in image, H _i' converge the tone value after calculating, H ₀for TINTC, k is step-length, works as H _i< H ₀time, k value, for just, is worked as H _i> H ₀time, k value is for negative.

In the process of seed selection and region growing, by Eigenvalue Criteria, to select sub pixel as the starting point of growth in image, then the pixel that sub pixel is around had to same or similar character with sub pixel in 8 neighborhoods merges in this region, again these new pixels are worked as new sub pixel and continued process above, until the pixel not satisfying condition can be included.Consider connectedness and adjacency between variant pixel, the Seed Points in area growth process under each searching loop is all dynamic change simultaneously.

To 8 neighborhoods of Seed Points periphery carry out region growing time, defined Rule of Region-growth:

|r-r ₀|≤k _r,|g-g ₀|≤k _g

Wherein, r ₀, g ₀selected any point P on image ₀r, G component, r, g are P ₀point is R, the G component value of the interior any point of 8 neighborhoods around, k _r, k _gfor corresponding to R, the selected threshold value of G component.

Experiment shows, the growth rule of definition has not only reduced computation complexity by this way, has accelerated sliced time, and has taken into full account the gradually changeable between each color component, is more conducive to the color characteristic of coloured image to extract.

In the region of image merges, it is close in color that the present invention defines two regions, adjacent on space, and its neighborhood place do not have significant edge be two can be connected region, region and the maximal value of the relative color distance in its neighborhood region are less than the threshold value of definition.The present invention adopts the color component average define color distance in region to calculate, and color distance is defined as follows:

$D_c=\frac{r_i\&CenterDot;r_j}{r_i+r_j}\left|\right|{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_i-{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_j\left|\right|$

R wherein _iand r _jrepresent respectively the number of pixels comprising in i and j region,

with

represent the color component average in two regions, || || represent Euclidean distance.R _ir _jproduct make to comprise the less region of number of pixels compare with the color distance in other regions little, thereby in the situation that color component average is identical, be conducive to the preferential merging of zonule, make segmentation result more meet people's visual characteristic.

In area growth process, if the color distance of selected seed region and its neighborhood is less than the distance threshold setting in advance, these two regions can be merged, after merging, again travel through each neighborhood, judge that again field is whether in threshold range, if be less than threshold value, intermittently carry out region merging, until do not have again phase near field to merge.

And in region, merge last, require:

If the number of pixels in a region is less than certain threshold value and color distance meets correlated condition, so this region is merged in the neighborhood region nearest with it and gone.

In the sub-block region to Tea Flower, choose in process simultaneously, likely owing to also there being close little region on old blade or petal, so if an area pixel number is less than certain threshold value, and during without adjacent merged region, by it removal.

Cut apart finally, then adopt Morphology Algorithm disposal route, to cutting apart rear image, expand respectively again and shrink process, to reduce little hole.Finally again the Tea Flower being partitioned into after completing is carried out to counting statistics.

Accompanying drawing explanation

Fig. 1 is that quick Tea Flower is cut apart and counting algorithm process flow diagram.

Fig. 2 is the original image of tealeaves.

Fig. 3 is the segmentation result of Tea Flower

Embodiment

Below in conjunction with accompanying drawing, illustrate the enforcement item with method of counting of cutting apart of quick Tea Flower.

The digital camera that experiment is used is CANON S80, in image capture process, adopt close shot pattern, close flashlamp, to avoid the impact of flashlamp self light on tealeaves color of image, should under natural light, carry out capture simultaneously, avoid the direct projection of sunlight, in capture, the imaging focal length of getting is about 30cm, and resolution adopts 1600 * 1200.

Experiment discovery, when capture focal length is nearer, obtained Tea Flower is less, and the accuracy rate of cutting apart is higher, and the accuracy rate of cutting apart when 10～20 Tea Flowers is the highest.When focal length is nearer, can affect the sharpness of the depth of field and imaging, thereby accuracy is cut apart in impact, but when focal length is farthest time, Tea Flower image reduces, the scope of image color and luster broadens, and the accuracy rate of cutting apart also can decline to some extent.

K _r, k _gfor corresponding to R, the selected threshold value of G component, when the threshold value of given R, G component and color distance, experiment is found, k _r=9, k _g=7, D _cvalue is 10 o'clock, obtains good result.

The original picture format of digital camera is rgb format, but there is very strong correlativity between R, G, B three-component, variation with illumination condition, R, G, tri-components of B are all more big changes, directly utilize these components often can not obtain required effect, so in the choosing of image color space, choose HSI space.

Under HSI model, the color information of image is mainly reflected by H and S, as follows to the conversion formula in HSI space from RGB:

$H=\left\{\begin{array}{cc}2\mathrm{\&pi;}-\mathrm{\&theta;}& B>G\\ \mathrm{\&theta;}& B\&le;G\end{array}\right.$

Wherein $\mathrm{\&theta;}=\arccos \left[\frac{[(R-G)+(R-B)]/2}{\sqrt{{(R-G)}^2+(R-B)(G-B)}}\right]$

$S=1-\frac{3\min (R,G,B)}{R+G+B}$

$I=\frac{1}{3}(R+G+B)$

Experimental result shows there is good connectedness cutting apart of this algorithm, experiment based on host CPU is being Intel Duo double-core I3-3220, on the computer of internal memory 4G, cut apart not enough 1S averaging time of piece image, thereby realized quickly and easily, Tea Flower is split from tealeaves image.For obtaining Tea Flower information fast, tea grower provides good method.But simultaneously owing to being Fast Segmentation based on color, have some tea tree petals owing to being dark green and can not be cut apart.

Claims (3)

1. the quick Tea Flower based on color is cut apart and a method of counting, it is characterized in that comprising following concrete steps:

(1) obtain the original image of tealeaves;

(2) by original image from RGB color space conversion to HSI color space, and the figure image intensifying that tone H wherein carries out based on feature tone is converged to calculating;

(3) image is converted back to RGB color space;

(4), according to the R of Tea Flower, G parameter, in image, select Partial Feature pixel as seed;

(5) based on growth rule, seed region is grown, the neighbor similar to seed color character is attached on the seed of growth district;

(6) based on merging rule and color distance, a plurality of sub-blocks region of entire image is scanned and chosen, to close in color, region adjacent on space merges;

(7) expand and the morphology that shrinks is processed in the region after being combined;

(8) complete cutting apart and counting of a plurality of Tea Flowers.

2. a kind of quick Tea Flower based on color according to claim 1 is cut apart and method of counting, it is characterized in that: the figure image intensifying of tone H being carried out based on feature tone in step (2) is converged in calculating, two Tea Flower feature TINTCs have been defined, near the tone value central value satisfying condition is converged to central value by different conditions and step-length, its step-length elects 1,3 as, 5 three kind different converges step-length, step-length near TINTC is little, more larger away from the step-length of TINTC, calculating is defined as:

H _i′＝H _i±k

1＜|H _i-H ₀|≤5 k＝±1

Wherein, 5 < | H _i-H ₀|≤10 k=± 3

10＜|H _i-H ₀|≤15 k＝±5

H _ifor the tone value of certain point in image, H _i' converge the tone value after calculating, H ₀for TINTC, k is step-length, works as H _i< H ₀time, k value, for just, is worked as H _i> H ₀time, k value is for negative.

3. a kind of quick Tea Flower based on color according to claim 1 is cut apart and method of counting, it is characterized in that: in the growth of step (5) seed region, defined Rule of Region-growth:

|r-r ₀|≤k _r,|g-g ₀|≤k _g

Wherein, r ₀, g ₀selected any point P on image ₀r, G component, r, g are P ₀point is R, the G component value of the interior any point of 8 neighborhoods around, k _r, k _gfor corresponding to R, the selected threshold value of G component.

Images